It is entirely possible that there is a correlation with both. The rainfall outside the growing period will influence soil moisture conditions during the growing season. Therefore, I would not be surprised to see a correlation with both parameters. While more parameters does not universally give a better model, it is certainly worth exploring the utility of all parameters for which a physical mechanism can be postulated.

In the case of rainfall, I can envision the relationship being non-linear. Too little or no rain is certainly bad, but I expect there is an ideal amount of rainfall and that production is negatively affected by too much rainfall as well.

There is always a good correlation between seasonal rainfall and productivity. However, seasonal soil moisture content (average) is better correlated than seasonal rainfall with productivity. The problem is to parameterize and run a model to assess soil moisture content. To carry the correlation try to make a data matrix of: year, productivity, seasonal rainfall. That is a matrix 3xn; where n = number of years. Plot in excell seasonal rainfall times productivity (x, y) see the tendency. For short data matrixes, linear relationships would be sufficient for larger data matrixes (n> 30 years) the data set tends to be non linear. Fit the tendency accordingly in the same excell program. Ask the program also for the equation and r2. Choose the model that describes best the tendency and use it for understanding if the next year seasonal rainfall is x, what would be the rice productivity.

Hope these comments would help you.

Kind Regards

Jose Navar, Ph.D

Thank you